Task-specific design of robotic manipulators

Abstract

Techniques for generating robot design specialized for a particular task are described herein. For example, a computer system can receive a first configuration of design parameters for a robotic manipulator that can be used to perform a particular task involving manipulating an object between a set of positions. The design parameters can include discrete hardware parameters. The computer system can set one or more constraints including a waypoint through which the robotic manipulator travels to perform the particular task. The computer system can generate, using an iterative algorithm and based at least in part on the first configuration of design parameters, (i) a second configuration of design parameters and (ii) a trajectory associated with the second configuration for performing the particular task between the set of positions by minimizing an objective function subject to the constraints.

What is claimed is: 
     
         1 . A computer system, comprising:
 a processing device; and   a non-transitory memory configured to store instructions that are executable by the processing device for causing the processing device to at least:
 receive a first configuration of design parameters for a robotic arm, the design parameters comprising one or more discrete hardware parameters and one or more continuous hardware parameters; 
 set one or more constraints comprising (i) a waypoint through which an end effector of the robotic arm travels to reach a second set of positions of a particular task and (ii) a set of points associated with an obstacle through which the end effector does not travel to reach the second set of positions, wherein the particular task involves moving an object using the end effector from any of a first set of positions to any of the second set of positions; and 
 generate, using an iterative algorithm and based at least in part on the first configuration of design parameters, (i) a second configuration of design parameters for the robotic arm and (ii) a trajectory associated with the second configuration for moving the object from any of the first set of positions to any of the second set of positions by minimizing an objective function subject to the one or more constraints, wherein the trajectory intersects the waypoint and does not intersect the set of points associated with the obstacle. 
   
     
     
         2 . The computer system of  claim 1 , wherein the objective function comprises a cycle time for the robotic arm to move the object from one of the first set of positions to one of the second set of positions along the trajectory. 
     
     
         3 . The computer system of  claim 1 , wherein the one or more discrete hardware parameters comprise at least one of a number of joints, a joint type, or an actuator size for the robotic arm, and wherein the one or more continuous hardware parameters comprise at least one of a segment length, a mounting angle, or a base location for the robotic arm. 
     
     
         4 . The computer system of  claim 1 , wherein the one or more constraints further comprise a torque limit for one or more joints for the robotic arm, and wherein the objective function comprises a torque on the one or more joints when performing the particular task or a maximum velocity of the one or more joints when performing the particular task. 
     
     
         5 . A computer-implemented method, comprising:
 receiving a first configuration of design parameters for a component of a robotic manipulator, wherein the design parameters comprise one or more discrete hardware parameters;   setting one or more constraints comprising a waypoint through which the robotic manipulator travels to reach a second position for a particular task involving manipulating an object between a set of positions including a first position and the second position; and   generating, using an iterative algorithm and based at least in part on the first configuration of design parameters, (i) a second configuration of design parameters for the component of the robotic manipulator and (ii) a trajectory associated with the second configuration for performing the particular task between the set of positions by minimizing an objective function subject to the one or more constraints.   
     
     
         6 . The computer-implemented method of  claim 5 , wherein generating, by the iterative algorithm, the second configuration of design parameters and the trajectory further comprises:
 generating a set of adjusted design parameters that satisfy the one or more constraints by iterating on the first configuration of design parameters;   determining a value for the objective function for each of the set of adjusted design parameters to generate a set of values; and   selecting the second configuration of design parameters by identifying an adjusted design parameter of the adjusted design parameters associated with a minimum value from the set of values.   
     
     
         7 . The computer-implemented method of  claim 5 , wherein the design parameters further comprise one or more continuous hardware parameters. 
     
     
         8 . The computer-implemented method of  claim 5 , wherein the one or more constraints further comprises a set of points associated with an obstacle through which the robotic manipulator is prevented from travelling, wherein the trajectory does not intersect with the set of points associated with the obstacle. 
     
     
         9 . The computer-implemented method of  claim 5 , wherein the one or more constraints comprise a torque limit for one or more joints of the robotic manipulator, and wherein the objective function comprises a sum of squared torques on the one or more joints when performing the particular task. 
     
     
         10 . The computer-implemented method of  claim 5 , wherein the design parameters comprise at least one of a type of joint comprising a revolute joint or a prismatic joint, a number of joints, a segment length between adjacent joints, or a joint mounting angle for the robotic manipulator. 
     
     
         11 . The computer-implemented method of  claim 5 , wherein the one or more constraints further comprise at least one of a maximum velocity and a maximum acceleration for the robotic manipulator performing the particular task along the trajectory. 
     
     
         12 . The computer-implemented method of  claim 5 , wherein the one or more constraints further comprise an end velocity and an end acceleration of zero at an end point of the trajectory. 
     
     
         13 . The computer-implemented method of  claim 5 , wherein generating the second configuration of design parameters further comprises:
 generating a set of adjusted design parameters by iterating on the first configuration of design parameters;   for each design parameter of the set of adjusted design parameters, determining a value of the objective function based on each combination of a start point of a range of start points in the set of positions and an end point of a range of end points in the set of positions for the robotic manipulator performing the particular task to generate a set of values of the objective function; and   selecting the second configuration of design parameters from the set of adjusted design parameters based on the set of values.   
     
     
         14 . The computer-implemented method of  claim 13 , wherein selecting the second configuration of design parameters from the set of adjusted design parameters further comprises:
 determining, for each design parameter of the set of adjusted design parameters, an average value of the set of values for the objective function; and   determining that the second configuration of design parameters is associated with a lowest average value for the objective function of the set of adjusted design parameters.   
     
     
         15 . The computer-implemented method of  claim 13 , wherein generating the second configuration of design parameters further comprises:
 for each design parameter of the set of adjusted design parameters, determining the value of the objective function based on each mass of a set of masses for the object to generate the set of values of the objective function; and   determining that the second configuration of design parameters is associated with a lowest average value for the objective function of the set of adjusted design parameters.   
     
     
         16 . The computer-implemented method of  claim 5 , wherein the iterative algorithm is configured to generate the second configuration of design parameters simultaneously with the trajectory. 
     
     
         17 . One or more non-transitory computer-readable media comprising computer-executable instructions that, when executed by one or more processing devices of a computer system, cause the computer system to perform operations comprising:
 receiving a first configuration of design parameters for a robotic manipulator, wherein the design parameters comprise one or more discrete hardware parameters;   setting one or more constraints comprising a waypoint through which the robotic manipulator travels to reach a second position for a particular task involving manipulating an object between a set of positions including a first position and the second position; and   generating, using an iterative algorithm and based at least in part on the first configuration of design parameters, (i) a second configuration of design parameters for the robotic manipulator and (ii) a trajectory associated with the second configuration for performing the particular task between the set of positions by minimizing an objective function subject to the one or more constraints.   
     
     
         18 . The one or more non-transitory computer-readable media of  claim 17 , wherein the objective function comprises a cycle time for the robotic manipulator to move the object from a start point to an end point along the trajectory. 
     
     
         19 . The one or more non-transitory computer-readable media of  claim 17 , wherein the one or more constraints further comprise a torque limit for one or more joints of the robotic manipulator, and wherein the objective function comprises a sum of squared torques on the one or more joints when performing the particular task. 
     
     
         20 . The one or more non-transitory computer-readable media of  claim 17 , wherein the one or more discrete parameters comprise at least one of a number of joints, a joint type, or an actuator size for the robotic manipulator.